As part of the Asset Management Council’s AMSPEAK virtual event, Darren Chuang and Hanbit Cho were asked to present on their ground-breaking work on redefining lifecycle degradation models. These concepts look to real-time nature of using available data to drive longer term decisions on managing assets.
The full webinar is linked below, but here’s a summary of the concepts covered.
PREDICTING FOR REPAIR
To efficiently plan for the maintenance, repair or replacement of assets, we need to know how they are degrading. To do this we use Lifecycle Degradation modelling which helps us predict what repair or replacement is needed and when.
Enhancing the predictability of building assets is a rising tide that lifts all boats. It enables line-of-sight across all stakeholders and empowers evidence based decision making.
To make a degradation model we need some essential ingredients:
How much is it going to cost to maintain, repair or replace an item?
How often do we need to maintain, repair or replace it?
When to action: when do we need to replace it?
Dynamic degradation estimation is redefining how often and when to action by making specific time points more accurate. Let’s look at the basic mechanics of an asset degradation model.
Consider a degradation model for a steel frame window in a coastal location, within a couple of kilometres of the nearest ocean. The graph shows the future condition scale, ranging from one as "new" to "unserviceable" as five. The gradient represents the rate of degradation and is calibrated annually, based on work orders or time-based information. This data is aggregated and fed back into a core model library.
Example of a degradation model
IoT AND ‘LIVE’ DEGRADATION MODELLING
Another example is a recent research project carried out in collaboration with the NSW Department of Education. The aim was to look at how real-time data from IoT devices can be used to create better degradation models. This involved using motion sensors to infer how much an asset is being utilised, and then using this information to better estimate the degradation of assets.
The project involved 26 motion sensors installed across two schools, in several different locations such as classrooms and corridors. Data was recorded over eight months, on the days when students were at school. This identified that some classrooms had significantly more students each day than others, as well as greater variability in terms of students per day.
Using mathematical modelling, the data is used to determine an accurate measure of asset utilisation, Duty Factor, as a factor of Relative Occupancy Rate and Relative Room Capacity. So if a carpet had a Duty Factor of two, this would indicate that it was used two times more than the average amount.
The equation is weighted differently for different industries. For a school, Relative Room Capacity is more significant because how much space students have in a classroom is a better indication of utilisation. Whereas in a shopping centre, Relative Occupancy gets a higher weight, because shopping centres are built to accommodate very large capacities and the number of visitors varies significantly from day-to-day.
By calculating asset utilisation, we can factor this into our degradation model across different time spans, making it dynamic and more accurate. We can then compare a specific asset against industry averages and see if an asset is degrading more rapidly at different periods of its lifecycle.
Using real-time IoT data enables more accurate models and makes it possible for decision makers to control the rate of degradation of assets. In the classroom example, this would be increasing the utilisation rate of the lowest used classroom and extending the life expectancy of assets in other classrooms.
Degradation modelling is constantly being refined and tested on a wide range of assets. We’re currently examining the impact of utilisation on external versus internal assets, and how different weather patterns can affect the rate of degradation.
Ultimately, lifecycle degradation modelling enables a proactive approach to maintenance, by predicting degradation, as well as preventing it or slowing it through the more efficient utilisation of assets. Increasing asset lifecycle is increasingly important for sustainability.